I met up with Noam from Mobixell before this year’s Mobile World Congress. I was meant to meet up with him again in Barcelona, but unfortunately he was ill.
Anyway, I enjoyed Mobixell’s take on the issue of network congestion – the causes of it (mainly video according to Mobixell) and some ways to try and fix it. I know they’ve had some success recently with operators, so I asked Noam to highlight some of the issues of congestion and what the operators should do about it.
Over to Noam:
– – – – –
It can’t be disputed that mobile video is a growing cause of network traffic with the rise in the popularity of “over the top” video – 60% of all HTTP traffic today and projected to reach 64% of all mobile traffic by 2013 – putting video optimisation firmly on most, if not all, operators’ agendas.
For mobile operators, the Holy Grail is to find the right balance between user experience, network availability and profitability. Let me illustrate the issue — imagine a group of teens hanging out:
One of them receives a link to an amusing video and starts to watch it. His mates want to watch so he starts the clip again. However they can’t all watch around the same phone so someone else gets their phone out to watch it. When the first clip is over they both browse other suggested videos. The first is boring, the second one’s slow to load, the third is good to start with but then it gets dull so they switch it off. The fourth video is great, but then the phone rings.
How many videos did they watch? How much of each one? And most importantly, how should an operator optimise those videos to get the best results for the consumers and the network?
In our recent “Viewer abandonment” studies we found that about 50% of viewers browse away from video clips after watching less than half. Knowing how users watch video and what they watch can make a dramatic difference to operators, saving network resources in the place of ‘optimise all’ strategies.
Whilst video optimisation can reduce video network traffic by as much as 40%, operators can keep the network resources required for optimisation manageable by employing user-aware video optimisation techniques. These continuously analyse user viewing patterns in real-time, deciding where best to implement optimisation resources. Operators can optimise only the most viewed videos, rather than every clip that crosses the network, thus making the desired bandwidth savings and also reducing their optimisation resource requirements.
User-aware video optimisation techniques include caching and optimising the most frequently viewed videos, or the most viewed sections of each clip, rather than the whole clip itself, and ‘just-in-time’ delivery, which serves each video stream at the same pace at which it is being viewed, opening up bandwidth for more concurrent sessions.
Both for individual and statistical purposes, it’s key to know how long users watch before browsing away, as well as how they pay and where they are when they stream video – at home or roaming, for example. Operators can choose whether to apply optimisation on an individual basis according to their rules databases, for example setting how optimisation is applied for video streamed to prepaid users who are charged by packet, or delivering only optimised video to roaming users by default, but giving them the controls to turn off optimisation.
In this manner, video optimisation techniques reduce the volume of data required to deliver video traffic across the network. Operators can defer expensive network upgrades while ensuring excellent user experience and what’s more, user-aware optimisation helps them to significantly reduce the amount of network resources required for implementing optimisation. For subscribers, user-aware optimisation gives them the best viewing experience possible, whether they choose to watch clips all the way through or whether they use their imagination to fill in the ending.
– – – – –
Thank you for this Noam. You raise a very valid set of points — why should operators spend a heck of a lot of effort optimising 100% of videos when the vast majority are abandoned so readily? With today’s handsets getting better and better, and — actually — cheaper and cheaper, mobile video optimisation is an issue that’s going to continue to cause network trauma.